An evaluation of canopy gaps in restoring wildlife habitat in second growth forests of Southeastern Alaska1

Paul Alaback

FINAL REPORT

February 20, 2010

1 A cooperative project with The Nature Conservancy-Alaska, Thorne Bay Ranger District and Craig Ranger District, Tongass National Forest, and POWTEC.

Executive Summary

We report on our initial findings from a two year study on a 20-year remeasurement of canopy gap treatments in second growth forests on Prince of Wales Island in Southeastern Alaska. Seventy-six gaps were selected for sampling representing a broad geographic, and ecological range of stand conditions throughout the region. Our analysis of these plots suggest that canopy gaps represent one of the most effective techniques for long-term improvement of habitat for deer and associated wildlife species in second growth forests on Prince of Wales. Our data shows statistically significant increases in species diversity, understory cover, forb biomass, and shrub annual growth for gap plots as compared to either thinned or unthinned controls. Canopy gap treatments create habitats that are on average 4 times the deer carrying capacity of our thinned second growth stands in the summer or over 8 times the carrying capacity of thinned sites in the winter. Within a gap there is as much summertime blueberry (Vaccinium) biomass as in typical old growth forests. A simple model (TONGASS GAP) was constructed for estimating the overall effect of canopy gaps on deer habitat at the stand level and also to provide managers with a tool to examinine tradeoffs between gap size, density and deer habitat for closed-canopy second growth stands. From this model we estimate as much as a four-fold increase in deer carrying capacity for winter habitats when up to 50% of a given stand has gap habitat. More typically gap treatments have created 5-10% gap habitat in clearcut units which should result in as much as a 50% increase in winter deer carrying capacity. It appears likely that these gaps will persist well into the future of these stands, since there was no significant increase in tree saplings following gap treatment, and there was no detectable influence of gap size on vegetation response. It will be highly desirable to continue monitoring canopy gap treatments to determine the overall longevities of these gaps, to determine the functional upper limit of gap size, and to determine the best ways to incorporate these treatments into overall stand and landscape management in the region.

Scientific and Management Context

One of the most difficult and long-standing conservation issues facing residents and land managers in Southeast Alaska has been the log-term impact of clearcut logging old-growth forests on wildlife habitat. Studies dating back to the 1960’s and 1970’s documented how logging has the potential to negatively effect many species of interest such as Sitka black- tailed deer, wolves, and goshawks in this region (TLMP planning documents, Wallmo and Schoen 1979, Alaback 1982, Hanley et al. 1985). Over 400,000 acres of highly productive, old-growth temperate rainforest in southeast Alaska have been logged since 1950. This timber harvest has been concentrated in the most productive and economically valuable forest stands at low elevations which have historically been important areas for people in local communities for hunting and subsistence, and as habitats for critical

2 wildlife species. Over half of the timber harvest has taken place on one island – Prince of Wales, in southern Southeast Alaska, since it has the largest concentration of easily accessible productive forest habitats in the region.

In Southeast Alaska there are many specific ecological factors which explain why logging can have such a negative impact on key wildlife species in this region. Most logging has occurred in low-elevation valley bottoms (<1000’) which provide critical habitat for wildlife, especially during times of heavy snow cover. Removal of old-growth forest and its replacement by second-growth forest affects winter habitat for deer in two specific ways: loss of snow shedding capability of complex old-growth canopies (effects mobility and foraging efficiency of deer) and loss of a productive understory plant community (provides forage quality and quantity). Although clearcut harvesting does produce an immediate flush of high quality understory biomass, it typically lasts only 10-25 years, and is not available to deer during the periods of heavy snow. The greatest impact occurs three or more decades after logging, during the “stem exclusion” phase of forest stand development, when the densely stocked and rapidly growing young conifers shade out most of the important plant species for deer and other wildlife species. The stem exclusion phase lasts for as much as 150– 200 years so can create a long-lasting deficit of wildlife habitat for a given watershed or region, unless an effective restoration strategy can be developed (Alaback 1982).

Over the succeeding 30 years since this issue of logging impacts has been understood scientifically, most efforts at restoration have centered on using techniques such as silvicultural thinning to stimulate understory vegetation growth (Hanley 2005, McClellan 2005). While thinning can be effective in improving wildlife habitat 5-10 years following treatment, one of the key limitations of this treatment is its relatively short longevity. This should not be too surprising since thinning is an agricultural technique dating back to at least the 1700’s designed to stimulate the growth of crop plants. By thinning a forest, the nutrients and other resources of a given forest are concentrated on a smaller number of trees resulting in increased growth rates and individual tree productivity as predicted by the -3/2 law (e.g. Oliver and Larson 1990). As a by-product, thinning can stimulate understory vegetation at first, since sunlight and nutrients become more available immediately after treatment. Soon, generally not more than 15 years, crop trees expand their branches and create a dense overstory canopy which shades out understory forage plants once again. While more intense thinning treatments (wider average tree spacing) may lengthen this process to a certain extent, data available to date suggests that on productive sites thinnings even up to a spacing of 20’ will still produce only short-term benefits to wildlife habitat (Alaback, unpublished data).

One relatively unstudied experimental treatment that appears to hold great merit for improving understory vegetation forage availability and diversity is the creation of artificial canopy gaps. Gaps represent small (<1/2 acre, or less than 160’ in diameter) clearings that simulate wind disturbance or small patch tree mortality characteristic of old

3 growth forests in Southeast Alaska (Juday and Ott 2002). Each gap is large enough to provide enough canopy opening and sunlight to produce significant forage, yet appears to be small enough to prevent a “conifer flush” typical of larger clearcuts or strip thinning prescriptions. Thus, the benefits of understory productivity are expected to last much longer than with conventional thinning. Moreover, understory flora within the gap includes evergreen forbs that represent a critical food source for a number of species in winter, as well as providing for snow interception and thermal cover along the edges. The forest structure created by canopy gaps would be expected to be more similar to the patchy forest conditions that characterize old-growth forests than what would result from any of the thinning treatments that have been studied.

Gap thinning applied at a landscape scale, combined with conventional thinning prescriptions, represents an innovation that may meet multiple objectives for conservation of biodiversity and timber management on clearcut landscapes. Only a fraction of the area generally requires a gap treatment (<5-10%), so the additional cost is small compared with the benefit of increased understory productivity and diversity within an otherwise unproductive landscape. Meanwhile, at the same time conventional thinning prescriptions will serve to meet timber management objectives within the larger landscape matrix. Within the managed landscape, this new regime moves beyond the artificial dichotomy of conservation versus development, and toward a more integrated approach where larger landscapes serve to meet multiple biodiversity and resource management objectives.

Nearly 600 gaps were installed on Prince of Wales Island from 1983-1993, and over the past few years many more have been installed across the region. The stated objectives are to “maintain forage production and habitat diversity for deer in stands 25 to 35 years old” and to “simulate old growth habitat conditions by opening up holes in the overstory to stimulate production of understory forbs and shrubs while also providing snow intercept”. While opinions of managers and residents vary as to the perceived value of these treatments, there has been no detailed scientific study of the effectiveness of canopy gap treatments in Alaska until we initiated this study in 2008. The Forest Service made two major efforts to monitor a sub-set of those gaps in 1990 and again in 1994 (Demeo 1990, Knotts and Brown 1995). Because of the efforts of the Forest Service to monitor and document these treatments, and the efforts of several forward-thinking managers to carefully archive records and data relating to these treatments we had the unique opportunity to examine the 20 year response of vegetation to canopy gap treatments in this study.

4 Objectives of this study:

General: To determine the overall effectiveness of canopy gaps in restoring productivity of understory vegetation and wildlife habitat to second growth stands on Prince of Wales Island.

Specifically:

1. Document long- term response of understory vegetation to canopy gap treatments on Prince of Wales Island. 2. Determine the influence of canopy gap size on: a)vegetation composition, b)forage quantity, and c)tree regeneration. 3. Determine long-term effects of thinning on vegetation response to canopy gaps. 4. Evaluate influence of site conditions on vegetation response 5. Develop decision support tools for the development of canopy gap prescriptions in second growth management.

Methods

Selecting Study Sites We did extensive surveys of over 100 canopy gaps (from the nearly 600 established across the region) to select sample sites for this study. The Forest Service provided extensive records including GIS data, study documentation and copies of field datasheets from the original study. From these data we identified potential study sites, which were then field verified, and ranked in terms of their suitability for this study using the following criteria:

1. Availability of retreatment and post-treatment monitoring data following methodology of Demeo (1990), and clear markings on the ground to allow accurate relocation of plots.

2. A well-distributed sample of canopy gaps across a broad range in diameter (from 40- 150’ in diameter).

3. An adequate sample of both thinned and unthinned stands.

4. Sites with well drained soils and productive plant associations (e.g. the western hemlock (100) and Sitka spruce series (300) of the Region 10 plant association classification (Demeo et al. 1992) so that meaningful comparisons could be made between canopy gap treatments and adjacent untreated areas2.

2 Note that it is common practice to establish canopy gaps in”natural openings” assuming that treatment of these openings will be a more cost-effective and long-lasting than when

5

5. Sites at low elevations (<1000’), on gentle slopes, and with similar soils and hydrological conditions.

6. Gaps need to be uniformly treated, e.g. all trees cut at same time in similar way. Avoid girdled plots where possible. 7. Gaps need to be treated symmetrically. All gaps should not have widths less than 1/2 the length of the gap or vice versa.

9. Gap and second growth transects should not be located within 100' of major openings such as a recent clearcut edge, road, wetland, or stream. Similarly we should not sample gaps that are within 100’ of mature forest (or the cut unit boundary) so that treatment response, in particular sunlight penetration through the canopy is not influenced by conditions outside of study plots.

After extensive surveys across the island we decided to focus on seven study areas so that we could represent a broad geographic and ecological range across the region, but also to allow examination of site-specific factors (Figure 1).

Table 1. Sampling matrix for canopy gap study. Each number represents the number of stands (gaps) sampled

Gap size Small <250 m2 Medium 250-600 m2 Large 600-1100 m2 Total Thinned - 16 14 30 Unthinned 19 10 17 46 Total 19 26 31 76

creating openings in uniform forest. Many of the gaps established in 1990 were in fact created this way. For purposes of this study these were not suitable since by definition this creates a contrast in soils, hydrology, and site conditions between the plots in the gap treatment and those in the adjacent control (untreated) plots, so that a scientifically meaningful comparison cannot be made.

6

Figure 1. Locations of study sites for canopy gap study. Numbers refer to specific study areas: 1=Polk Inlet, 2= Rock Creek, 3=Staney Creek,4=Deer Creek,5=Sandy Beach, 6=Luck Lake, 7=Twin Island. (Adapted from Mike Ausman graphic).

7 SECOND GROWTH TRANSECT GAP TRANSECT PLOT 10 ft

EDGE EDGE 5-17 ft S8 S6 S4 S2 G0 G2 G4 G6 G8

S7 S5 S3 S1 G1 G3 G5 G7

REBAR OR PVC STAKES MAIN GAP STAKE -- WITH UNIQUE TAG NUMBER FISHEYE PHOTO LOCATIONS Figure 2. General plot design, after Demeo 1990. Note that plot spacing in gap plots varies depending on the diameter of the gap.

Field Measurements

We followed the methods of Demeo et al. (1990) in sampling trees and understory vegetation as described below. Plots were originally marked with pvc or rebars at the gap edge (plot E0) and at the end of the gap and second growth transects. Wherever possible transects were established in precisely the same spots that were used for monitoring in the past. Sometimes this required re-establishing transects and re-marking each end to facilitate future remeasurement. We now have accurate GPS locations for all sites that were sampled (and all that have been surveyed as potential study sites for next year).

Eight plots were established, equally spaced across canopy gaps, and one plot was established at the edge of the canopy gap, and 8 plots were established in the adjacent forest using 10’ spacing starting at the edge of the gap. Each plot was 1x2 meters in size and was sampled with a portable ½” pvc plot frame. The length and width of the gap was measured to compare gaps of different sizes and shapes and to evaluate changes in gap size over time. We also took photo points with a digital camera looking across the gap (G0 to G8) and across the second growth transect (S1 to S8) to document general vegetation patterns, as was done in the original study. All photos are included with the data in our databases, and the databases include references to photo file names so that it will be easy to make comparisons of these photos with those taken in the future3

3 Unfortunately, while much effort was originally devoted to taking precise photographs of study plots in exactly in the same way as we did in this study, these original slides appear to have been lost. If they are ever located they would provide an invaluable resource in documenting how vegetation has changed over time in these plots and to determine how site specific conditions may influence vegetation response.

8

Tree measurements Percentage cover of all overstory trees were estimated as well as the percent cover of each overstory species. Average height (in feet) of the overstory and of each tree species was estimated with clinometer readings of representative trees, and estimations of adjacent trees. Trees were tallied by species which were greater than 0.4” (1 cm) in basal stem diameter. Breast height diameters were measured of all trees greater than 1” DBH that were rooted in the plot, as in the original study.

A. B.

Figure 3. Examples of fisheye photos of overstory canopy used to estimate radiation interception by foliage. A)Canopy gap center (>35% radiation interception) B)Untreated control plot (<5% interception).

A new measurement for this study was to document the actual proportion of sunlight that is penetrating through the forest canopy. This is a key variable since it would be expected to explain how trees and understory plants respond to both local and stand level changes in forest structure, and also provides for a precise documentation of overstory structure. We took a series of 180 degree fisheye lens photographs from the three equidistant locations across the gap and second growth transects with the camera always facing magnetic north. We used a Nikon Coolpix 950 and 5000 series digital camera with a Nikon Fisheye add-on Lens. Camera sensor resolution was 2-3 MP. Since the photos were taken at known compass directions, and were exactly level with the horizon we were able to register these photos geographically, and to calculate diffuse and direct radiation penetrating through the forest canopy at different times of the day and the year using the GLA software developed by ecologists at the institute of ecosystem studies (http://www.ecostudies.org/gla/).

9

Vegetation measurements The percent cover and average height (feet) of all shrubs were estimated by species. For Vaccinium ovalifolium and V. alaskaense we counted the number of stems 0.4” (1cm) in basal diameter or greater and their average height. For this species group we also determined “% forage” by estimating the projected leaf area of leaves 1.2 meters (4.5’) above the ground or less. These data were also used to estimate forage biomass (annual leaves and twigs) using regressions derived from our clipping data. Lastly we recorded percent browse for randomly selected branches of these Vaccinium species. This was done by randomly selecting 21 twigs at the end of 3 or more branches and counting the proportion of twigs with evidence of browsing.

For ground vegetation we estimated % cover by species. Grasses, rushes and sedges were recorded as a group. Ferns were also recorded by species. We also estimated the percent cover of slash, and its average depth, and other major surfaces of plots including percent bare ground, percent stumps, and percent recent logs. While slash was a key concern in the original study, for this study all slash was well decomposed and incorporated to the forest floor, with exception of larger woody material such as larger tree bole segments.

Forage Biomass For this study we collected additional data on forage biomass so that we could more precisely document the wildlife habitat value of understory vegetation in sample plots. We randomly placed a 1x2m plot in each canopy gap and in the center of each second growth transect. We made the same measurements on Vaccinium and the other 4 key forage species as we made in the other 1x2m plots (percent cover, percent forage, percent browse, number of stems greater than 0.4” (1 cm), average height). We then clipped each species on the 1x2m plot. Subsamples were used to estimate the proportion of leaves and twigs, and to determine wet/dry weight ratios. These data were used to develop regression equations to estimate forage biomass for each treatment in each stand.

Tree talley plots To document general stand structure a sample plot was established in the center of each second growth transect, similar to what was done when plots were originally monitored, except that we used larger plots to account for changes in tree size and density. Plots 30’ x 30’ or larger were established in the center each second growth vegetation plot. In each sample plot the DBH, canopy position and species of all trees greater than 1” basal stem diameter were recorded.

Data Archiving and lab procedures Data were recorded onto excel spreadsheets at the end of the week or at the end of each field season. Biomass samples were oven dried at 60 degrees C for 48 hours and weighed to the nearest 0.1 gram. Weight data was then recorded on final spreadsheets of data for each plot. All of the spreadsheets were designed so that they had the same format and appearance as the paper field sheets, to minimize errors in entering data onto the computer. One spreadsheet (with three separate worksheets) was created for each sample

10 stand from a master template, and labeled according to gap and stand numbers. Excel Macros using VisualBasic, were developed to extract data from these worksheets into a series of data tables that could be uploaded into a database program for data verification and analysis. In some cases fairly complex macros were needed to account for variations in formatting of field sheets, and to simultaneously update multiple database tables at a time. Macros are available to facilitate processing of future data when these plots are monitored again.

Data Analysis Many kinds of data were collected in the field on each of the datasheets for each sampled stand. A database structure was designed so that data collected for individual plants, plots, treatments or stands could be easily integrated together for analysis. Separate tables were developed for the clipping biomass data, understory cover data, site variables, GIS derived site data, fisheye photo data, overstory and tree variables, second growth tree talley plots, and tallies of larger trees in the gap plots. Filemaker pro database software was used to manage the data and to produce summaries of data by plot and by treatment.

Data validation was performed by running statistical summaries of all variables by stand, treatment and by plot, and by using table lookup functions and other cross referencing with the database. Outliers were identified and cross checked with original datasheets to identify errors in data entry or in field procedures. Unknown species were verified and corrected after studying sample species and photos of these specimens.

Since the plot design provides a close pairing between second growth and canopy gap transects we determined statistical differences by running a Wilcoxon paired test. This test is quite conservative in that it is a non-parametric test so does not make any assumptions about statistical distributions. In most cases we have summarized statistical patterns between the thinned, unthinned, and gap-treated plots by plotting a boxplot, since this is the most parsimonious way to characterize patterns in data of this kind. In these plots the median value is represented by a horizontal line through the box, and the area of the box shows the distribution of 66% of the observations, and the error bars represent the range of 95% of the observations, and any observations beyond this range are indicated by a dot or circle (DalGaard 2002). Comparisons of more than two study areas or treatments were made using a one-way ANOVA if data were normally distributed or otherwise with a Kruskal-Wallis test. All statistical calculations were made with the statistical programming language R (a newer version of the S language, see Dalgaard 2002).

11

Figure 4. Alaska blueberry (Vaccinium ovalifolium-alaskaense) biomass estimation regression equations from plot measurements of percent cover. N=56 for gap plots, and 31 for untreated (including thinned plots). No significant difference was detected between gap and untreated plots.

Figure 5. Forb biomass estimation equation for forb species important to deer habitat (Tiarella trifoliata, Coptis aspleniifolia, Cornus canadensis, and Rubus pedatus). No significant differences were detected between gap and untreated plots so a pooled equation was used.

12

Biomass Estimation Forage biomass was estimated by multiplying wet weight to dry weight ratios from subsamples on all fresh weights for field samples on a given day. In a couple of instances wet/dry ratios were not available, so average estimates over the field season were used. Subsamples of leaf to twig ratios were used to estimate twig and leaf biomass for each clipping plot. Equations for estimating biomass were created by running regression analyses between dry weight data from clipped plots and corresponding data on plant cover, height, forage or stem density. While there were significant correlations between forage biomass and various combinations of stem height, density, or % forage, the most simple measure, % cover proved to be the most statistically reliable. When plots of data suggested a non-linear relationship (which most did) data was transformed to a natural log (ln) to linearize the pattern. Clipping plot data included the full range of Vaccinium cover class, although as in the plot data the majority of samples were less than 50% cover. The log-log regressions underestimated biomass when cover exceeded 50%, so our estimates of Vaccinium biomass are overall somewhat conservative. We did not detect any significant change in the proportion of twig and foliage biomass with respect to plant size or cover class, so a pooled estimate of 60% foliage and 40% twigs was used to estimate plot biomass.

There were few forb biomass samples so data for these species (Tiarella trifoliata, Rubus pedatus, Coptis aspleniifolia and Cornus canadensis) were pooled together which usually provides for more accurate estimates (see Alaback 1986, 1987). In general these biomass prediction equations account for about 70% of the biomass sample variation (Figures 2,3). Note that while precise relationships were found between % cover and biomass for these plots, these estimators are unlikely to provide accurate estimates for stands of other ages or stand conditions since cover does not account for structural changes in vegetation which would occur in stands of different ages. This is why it is key to collect clipping data on a subset of plots each time a series of plots is monitored (see Alaback 1986).

Site Productivity There is a wide range of tree productivity across Prince of Wales and Southeast Alaska in general which can have an important influence on vegetation responses to treatments (e.g. Alaback 1982, Hanley and Brady 1997). For this study site productivity was characterized several ways. For all study sites GIS data, and field data was used to estimate slope, aspect, soils type and elevation of all treatments. In addition we keyed out plant associations, according to the system developed for the Ketchikan Area by the R10 ecology program for all transects (Demeo et al. 1992). In addition tree site productivity was assessed by comparing average tree heights by species between study sites. Since most sites were the same age (45 years) this provided the most simple and straightforward metric. In addition we calculated site index for Sitka spruce and western hemlock trees on all sites using the equations published by Farr (1984). We were able to calculate site index by using our measurements of dominant or co-dominant tree heights taken along our transects. When three or more trees were measured for a given site we calculated site index assuming that Sitka spruce trees took 5 years to reach DBH height, or that western hemlocks reached this height at approximately an age of 10 years. Our

13 estimates of site therefore may over estimate some of the poorer sites where trees took more time to reach DBH height.

Deer Habitat Evaluation To provide a rough estimate of deer habitat value of understory vegetation in our sample plots we estimated the biomass of all critical deer forage species with a overall percent cover value greater than 0.01% using our biomass regression equations developed for Vaccinium canopy cover to other woody shrub species, and for all forbs from our combined equation for Tiarella trifoliata, Rubus pedatus, Coptis aspleniifolia and Cornus canadensis. These data were then used to estimate potential maximum deer carrying capacity using the deer habitat model FRESH (http://cervid.uaa.alaska.edu/deer/, Hanley and Rogers 1989, Hanley et al. 1989). We selected energetic requirements for summer using does with one fawn, so that we could assess the ability of these treatments to provide habitat for a stable deer population, and also for adults in the winter which is generally considered the critical time period for determining overall population size. Note that our calculations are for the potential maximum deer populations with these estimates. No consideration was made for forage becoming unavailable due to snow, or to other factors that could constrain deer populations such as wolves, extreme weather events, human activity, or landscape factors such as fragmentation and the existence of other habitats within the home range of a given animal.

Results

In the 76 gap treatments, selected from seven geographical regions within Prince of Wales we found surprising consistency in vegetation response to treatments. Highly significant (p < 0.001) differences between either thinned or unthinned forests and canopy gap treatments were found for all basic measures of plant diversity, abundance and structure (Figs 6 - 8, Table 2).

Effects on Plant Diversity Highly significant differences were found in overall vascular plant species diversity on gap plots as compared with either thinned or unthinned control plots. Overall 47 species were found in the canopy gap treatments as contrasted with only 28 in the control plots. Of the 50 species encountered on our plots 41 species were more common on gap treatments than control plots, and only two species were more common in the dense canopies of the control plots (Moneses uniflora and Streptopus streptopoides). There was little difference in the diversity of thinned vs. unthinned plots (20 vs. 28 species). Treatments were more similar in the number of common species (species greater than 0.1%, gaps=10, controls=6). The gap treatments had many more rare species. But overall there was little difference in species equitability (relative cover) between treatments (Fig 8).

14 Table 2. Understory plant species and their abundance 20 years after canopy gap treatments.

Treatment EDGE GAP UNTREATED OVERALL

Species Cover,% Freq. Cover,% Freq. Cover,% Freq. Cover,% RANK Herbaceous species Actea rubra 0.00 0% 0.00 1% 0.00 0% 0.00 36 Adiantum pedatum 0.00 0% 0.00 1% 0.00 0% 0.00 Athyrium filix-femina 1.27 16% 5.41 68% 0.80 42% 2.49 4 Blechnum spicant 0.09 4% 0.89 50% 0.27 45% 0.41 13 Cardamine oligosperma 0.00 0% 0.00 1% 0.00 0% 0.00 Circaea alpina 0.00 0% 0.05 5% 0.00 0% 0.02 26 Clintonia uniflora 0.01 3% 0.00 1% 0.00 1% 0.00 29 Coptis aspleniifolia 0.05 8% 0.28 26% 0.06 16% 0.13 18 Cornus canadensis 0.47 29% 1.12 67% 0.26 38% 0.62 12 Dryopteris austriaca 1.60 53% 3.04 97% 1.23 91% 1.96 6 Epilobium angustifolium 0.00 0% 0.09 3% 0.00 0% 0.03 22 Epilobium ciliatum 0.00 0% 0.00 4% 0.00 0% 0.00 34 Equisetum arvense 0.00 0% 0.00 1% 0.00 0% 0.00 37 Fauria crista-galli 0.01 1% 0.00 3% 0.00 0% 0.01 28 Galium spp. 0.00 0% 0.00 1% 0.00 0% 0.00 40 Galium trifidum 0.00 0% 0.02 11% 0.00 1% 0.01 27 Gaultheria shallon 0.00 0% 0.00 3% 0.00 0% 0.00 37 Gymnocarpium dryopteris 1.91 37% 2.15 80% 1.19 67% 1.75 7 Linnaeus borealis 0.00 0% 0.00 1% 0.00 0% 0.00 40 Luzula parviflora 0.00 0% 0.01 3% 0.00 0% 0.00 33 Lysichiton americanum 2.63 12% 2.85 37% 1.22 21% 2.23 5 Maianthemum dilatatum 0.03 4% 0.03 11% 0.01 5% 0.02 24 Moneses uniflora 0.00 0% 0.00 0% 0.01 8% 0.00 30 Polystichum munitum 0.01 1% 0.12 11% 0.04 8% 0.06 21 Prenanthes alata 0.00 0% 0.00 3% 0.00 0% 0.00 37 Ranunculus occidentalis 0.00 0% 0.00 1% 0.00 0% 0.00 40 Viola glabella 0.05 3% 0.01 7% 0.01 5% 0.02 23

15

Table 2. Continued. Treatment EDGE GAP UNTREATED OVERALL

Cover, Species Cover,% Freq. Cover,% Freq. Cover,% Freq. % RANK

Woody Shrub species Menziesia ferruginea 5.42 32% 8.29 79% 2.27 51% 5.33 3 Oplopanax horridum 1.51 4% 2.34 36% 0.41 14% 1.42 8 Ribes bracteosum 0.00 0% 2.12 14% 0.00 0% 0.71 10 Ribes lacustre 0.00 0% 0.01 1% 0.00 1% 0.00 31 Ribes laxiflorum 0.00 0% 0.90 38% 0.02 1% 0.31 15 Ribes spp. 0.00 0% 0.00 1% 0.00 0% 0.00 40 Rubus parviflorus 0.00 0% 0.05 4% 0.00 0% 0.02 25 Rubus pedatus 0.16 17% 0.91 63% 0.12 29% 0.40 14 Rubus spectabilis 3.63 33% 14.22 92% 1.09 47% 6.32 2 Salix scouleriana 1.32 1% 0.63 8% 0.00 3% 0.65 11 Sambucus racemosa 1.39 2% 1.00 12% 0.00 5% 0.80 9 Streptopus amplexifolius 0.00 0% 0.00 4% 0.00 3% 0.00 34 Streptopus streptopoides 0.00 0% 0.00 1% 0.01 7% 0.00 31 Thelypteris phegopteris 0.02 3% 0.18 11% 0.09 8% 0.10 19 Tiarella trifoliata 0.20 13% 0.47 50% 0.14 37% 0.27 16 Vaccinium caespitosum 0.39 1% 0.00 1% 0.00 0% 0.13 17 Vaccinium ovalifoium- alaskaense 17.99 76% 26.44 93% 7.58 89% 17.33 1 Vaccinium parvifolium 0.00 0% 0.22 3% 0.02 1% 0.08 20 Total plant cover, % 38.75 72.88 16.88 43.63 Plant species richness * 47 28 47 Species preferring gaps 41 Species preferring untreated 2 *Cannot estimate richness for edge plots because of difference in sample size (one plot per stand as contrasted with 8 plots per stand for gaps and thinned or untreated.

Effects on Vegetation Structure The most dramatic effects of canopy gap treatments as seen 20 years after treatment was the contrast in overall stature, density and structure of understory vegetation. Canopy gap treatments averaged over 58% plant cover as contrasted with 19% in control plots (Fig 6). For most wildlife species structure is the key characteristic of vegetation that defines the quality of habitat, so this is an important result. The critical deer browse, Vaccinium averaged 23% cover (as contrasted with 9% in the control plots). Two critical forbs for deer, Cornus canadensis and Rubus pedatus were 3 times or more common on canopy gap plots than on the control plots. There was no overall statistically significant effect of thinning on vegetation structure. Vaccinium averaged 9.1% cover on thinned plots, and 7.9% on unthinned plots for example. There was only a 1.5% difference in average cover between these treatments. This is consistent with previous studies in Southeast Alaska that

16 documented only transient effects of thinning treatments on wildlife habitat (e.g. Alaback 1984, Alaback, unpublished data).

Figure 6. Understory vegetation response to 75 canopy gap treatments. All differences are significant at P <.0001. Richness was calculated for all 8 2-m2 plots for each treatment block. Species richness for edge plots could not be calculated due to differences in sample size.

Figure 7. Understory cover 20 years following canopy gap treatments on 75 study sites. Differences between gaps and control plots significant at P < 0.001,

17

Figure 8. Rank-order relationships for understory species by treatment. The length of each line represents species richness, and the slope evenness. No significant difference in evenness was detected.

Understory Biomass We also detected highly significant differences in understory biomass between treatments for Vaccinium, the group of preferred forbs for deer (Cornus, Coptis, Rubus pedatus and Tiarella), as well as for overall shrub and herb biomass (Figure 9). Other measures of forage abundance such as density of Vaccinium stems were also significantly different between treatments. The one exception is percent browse. While high levels of utilization were evident in the canopy gap treatments, in the control plots browse utilization was highly variable, so no consistent difference was detected. There was no significant difference in Vaccinium height between treatments. This implies that shrubs are slowly dying back in the second-growth, and also suggests that our estimates of the magnitude of the differences in forage biomass between treatments are quite conservative.

18

Figure 9. Estimated understory biomass on gap and control treatments, based on regression estimates from clipping data and vegetation structure from other plots. Gaps are significantly different from control treatments at P < 0.001.

Effects on Deer Carrying Capacity To better assess the implications of these changes in vegetation composition and structure these biomass data can be used to estimate maximum potential deer carrying capacity. This analysis assumes our plots represent habitat conditions that may extend over an entire stand or the home range of a deer. The FRESH model developed by Tom Hanley and colleagues was used for this analysis (Hanley and Rogers 1989, Hanley et al. 1989). For simplicity this model was used to assess the potential maximum carrying capacity of deer, not considering effects of snow, predation or other factors known to constrain these populations. First these data were used to look at deer carrying capacity in the summer when nutritional demand would be the greatest especially for does with fawns (Fig. 9). Gap treatments actually resulted in levels of Vaccinium biomass similar to or even exceeding typical levels in old growth forests. Available forage biomass and utilization is similar for both old growth and canopy gap plots. The largest difference between these habitats is actually in the abundance of forb species. Even though forbs significantly responded to gap treatments, generally with a three fold or more increase as compared to untreated second growth plots, they still were less than half what typically occurs in an old growth forest habitat. This explains why in terms of deer carrying capacity (or deer days per ha of habitat) old growth habitats still have significantly greater quality than any of the second growth habitats. Second growth plots without canopy gaps generally have approximately 17% the carrying capacity of old growth plots. With canopy gap treatments carrying capacity averages 72% of old growth sites.

19

Figure 9. Changes in deer carrying capacity and utilization of understory vegetation resulting from canopy gap treatments for 76 study sites on Prince of Wales Island. Results from the model FRESH assuming summer climatic conditions and does with one fawn (Hanley et al. 1989). Old growth data from Hanley and Brady (1997).

It is generally thought that for deer and many other species winter is the most critical period for looking at carrying capacities of a given habitat. In this case there is a much more dramatic fall off in habitat quality in second growth habitats as compared to averages for old growth habitats. Gap treatments increased the carrying capacity for deer on second growth sites over eight times. Even so, the carrying capacity is still less than a quarter that of a typical old growth habitat.

20

Figure 10. Winter carrying capacity for deer 20 years after canopy gap treatments on 75 sites on Prince of Wales Island. Old growth data from Hanley and Brady (1997).

Overstory Effects Overstory canopy cover varied widely across most gap transects, from complete cover near each end to mostly open in the center. This may have become more of an issue over time as some of the gaps have begun to fill in. One of the biggest surprises of the study has been the lack of a strong pattern with gap size. No significant differences were detected in plant species richness, biomass or cover, or tree seedling density, between gaps of small, medium or large sizes (Figure 11). This is even more surprising considering that gap areas varied from 60 to over 1000 m2 (table 1). Several gaps have filled in significantly since they were established 20 years ago, but again we were unable to detect an effect of gap size in influencing this pattern. Similarly we had expected higher rates of tree colonization and growth in the larger gaps. Tree regeneration in gaps was highly variable. Saplings on most sites declined significantly compared to control plots. Seedlings however are nearly as abundant on gap plots as they are on unthinned controls (Table 3). Even overstory cover was not significantly different between treatments, mostly due to high variability in the gap plots. I suspect that part of the issue here is variability in how gaps were originally created. For example variation in retention of mid sized trees could easily confound patterns of tree colonization. There also could be many factors that influence the speed with which canopy gaps fill in by surrounding trees. The fact that there is no clear pattern of tree regeneration with canopy gap size, and even after 20 years the greatest increases are in seedlings, not saplings may suggest these gaps will persist long into the future.

Surprisingly few strong correlations were found between measures of overstory canopy structure and vegetation response. This might in part be explained by the complex nature

21 of the canopy in gap plots, 20 years after treatment, and also a possible lag-effect in vegetation response to changes in canopy conditions.

Figure 11. Effects of gap size on understory vegetation (a) and tree seedling (b) from 75 sites on Prince of Wales.

Canopy gap treatment resulted in large increases in radiation interception, even 20 years after treatment. Approximately 10 to as much as 70% radiation penetrates most canopy gaps, as contrasted with less than 5% radiation penetration in most control plots (Fig 12). Gaps in thinned stands generally had higher radiation levels, but we were unable to detect any vegetation response to this subtle increase in radiation (Fig 13).

Figure 12. Effects of canopy gap size on solar radiation penetration through forest canopies.

22

Table. 3. Characteristics of Sampled Stands GAP THINNED UNTHINNED Variable Units Mean SE Mean SE Mean SE Overstory Hemlock Site Index m/50 years 13.1 (2.0) 14.2 (3.0) Spruce Site Index m/50 years 17.1 (2.4) 17.3 (3.7) Basal Area m2/ha 54.4 (23.5) 61.8 (35.2) Cedar m2/ha 0.1 (0.5) 2.3 (8.9) Hemlock m2/ha 27.4 (20.5) 37.3 (24.5) Spruce m2/ha 24.6 (15.9) 20.2 (24.8) Density Trees/ha 852.8 (408.8) 1864.7 (1149.3) Mean Diameter (QMD) cm 25.0 (7.8) 15.6 (7.3) Tree regeneration Seedling Cover % 25.6 (20.7) 10.6 (10.9) 17.6 (18.9) Sapling Cover % 58.8 (60.7) 87.2 (10.8) 76.2 (16.1) Seedling height m 2.8 (2.9) 1.2 (1.0) 4.1 (3.9) Sapling height m 13.1 (3.4) 13.6 (1.9) 13.7 (3.8) Seeding density Trees/ha 8474 (44837) 625 (741) 10109 (37613) Sapling density Trees/ha 1410 (5564) 781 (809) 16168 (52682) Spruce seedlings % 0.1 (0.2) 0.0 (0.1) 0.1 (0.1) Spruce saplings % 0.3 (0.4) 0.3 (0.2) 0.3 (0.2) Gap Characteristics Gap Length m 20.8 (7.2) Gap Area m2 509 (302) Gap Age Years 20.6 (1.1) Radiation Interception Diffuse % 35.3 (13.3) Direct % 27.9 (15.3) Total % 31.6 (14.0) Stand Characteristics Age of thinning Years 27.0 (4.4) Stand Age Years 44.4 (2.1) 42.3 (4.7) Slope % 14.5 (11.5) 13.5 (11.5) Aspect Degrees 151.7 (101.4) 120.8 (76.0) Elevation m 162.7 (67.2) 34.7 (40.2) Number of Stands Sampled 75 29 46

23

Figure 13. Effects of thinning on vegetation response 20 years after treatment. No effect could be detected in either control or gap treated plots.

Figure 14. Shrub cover contrasts between study areas.

Site Effects Since there were not strong or consistent correlations between gap size or radiation penetration and vegetation response it would be expected that site effects (e.g. nutrient availability, or climatic differences) may be playing a key role in determining vegetation response to treatments. Surprisingly there were not strong consistent differences in overall cover or biomass that directly corresponded to study area. Twin Island as a karst area did show some differences when compared to other sites. Shrub cover was high for these sites, which would make sense since they are highly productive sites. Vaccinium biomass on the hand was lower on these sites, primarily because of the dominance of

24 species such as salmonberry (Rubus spectablis). Rock Creek had relatively low forage values for wildlife, so was distinctive from other sites. Other measures of site productivity such as Sitka spruce site index, hemlock site index, and overall tree height did not have significant correlations with vegetation response to gaps. Substantial variation did occur with respect to plant associations and soils types, so these may warrant further investigation.

Side Lighting Another question of interest for resource management is the extent to which gap treatments influence the surrounding forest. It is often thought, for example, that other thinning techniques such as low-cost strip thins may have similar benefits to canopy gaps by providing extensive side lighting in adjacent stands. In this study we were unable to detect significant effects of side lighting. As mentioned above there was not a detectable influence on vegetation in thinned vs. unthinned stands, even though we could measure significant increases in solar radiation in these sites. We also examined the effect of side lighting by looking at overall vegetation responses according to plot positions from canopy gap treatments (Figure 15). Plot G0 which is located right on the edge of the gap generally had vegetation responses intermediate between the gap plots and control plots. However plot S1, which is only 10 feet away from the gap edge had no detectable influence from the gap, and we could not detect any changes in vegetation within the control plots that related to side lighting from the gap treatments. Likewise we were unable to detect any consistent changes in vegetation response within the gap treatments.

Figure 15. Effects of distance from gap treatment on vegetation response. Error bars represent the standard error of estimates of mean shrub cover for all plots at a given distance along the transect (n=76).

25

Management Recommendations Our study results suggest that canopy gap treatments may be an important treatment to consider over a broad range of site conditions in restoring the wildlife habitat and plant diversity in second growth closed-canopy forests on Prince of Wales Island. Over a broad range of soil types, site indices, plant associations, and locations across Prince of Wales we detected a remarkably consistent and strong vegetation response to canopy gap treatments, even 20 years after the treatment was installed. As suggested in previous studies thinning has a relatively short life-span in terms of its ability to stimulate understory growth (Alaback 1984, Alaback unpublished data and unpublished reports). Data from research plots suggested a 12-15 year life span for thinning effects on understory plants. The operational scale thinnings studied here appear to confirm these earlier findings with little detectable vegetation response 20 years after treatment. The fact that canopy gaps appear to provide a much longer lasting effect in improving wildlife habitat values in these second growth forests is a key consideration in giving more emphasis to this treatment where long-lasting results are desired.

Figure 16. Results of TONGASS GAP calculator on predicting effects of canopy gap treatments on seasonal deer carrying capacity for thinned second growth forest. A hypothetical 100 acre 45 year old second growth stand would be expected to increase its carrying capacity for deer either in winter or summer based on the average spacing of a 60’ gaps (left) and by the number of gaps created in the unit (right). It is assumed the canopy gap treatments were made at the same time the unit was thinned at stand age 20.

To properly evaluate canopy gaps as a restoration treatment, and to determine how to best apply it to any given watershed or stand it is necessary to understand the overall effects of this treatment on a whole stand, not just the gaps themselves. For this purpose we developed a simple spreadsheet model TONGASS GAP derived from our estimates of deer carrying capacities presented in figs 9-10 (See appendix I for details). For typical gap treatments, where as little as 5-10% of the area of the stand is treated, we estimate there will still be a 20-50%. increase in deer carrying capacity. In theory as much as a 4 fold increase in deer carrying capacity could be achieved in the winter, or a doubling of summer carrying capacity if canopy gaps were increased to 50% of the stand area. In addition to these significant gains in habitat quality, canopy gaps would be expected to

26 also be an important means to promote connectivity, dispersal habitat and to retain pockets of understory diversity that could aid reestablishment of diversity when stands are scheduled for other treatments such as commercial thinning. Clearly there is much promise in using canopy gaps for restoration treatments, and we would recommend that land managers give serious thought to this prescription following clearcutting where it is practical to do so on Prince of Wales Island, and for similar sites elsewhere in Southeast Alaska.

How long canopy gaps can provide an important habitat function in the life of a second growth forest is still an unanswered question. Limited tree regeneration in many gaps suggests that they should persist for at least another decade. Continued monitoring will be needed to determine how long these treatment effects will persist. Even if these gaps only persist another 10-20 years, this should be well within the time span that commercial thinning can start to be applied to these stands. It will also be important to understand how slope and aspect may directly or indirectly effect vegetation response to these treatments.

From an ecological standpoint there is much data and theory that supports the idea that forest biodiversity is generally enhanced by increasing forest heterogeneity as we have done with the creation of canopy gaps (e.g. Hanley et al. 1989, Pickett and White 1985, Veblen and Alaback 1995, Watt 1947). It makes sense that disturbances which create irregular openings are generally going to create a variety of ecological conditions which will provide habitat for a wider range of species than what would occur with more homogeneous forest conditions (or more homogeneous disturbances). There is considerable evidence that canopy gap formation is a major driver of ecological diversity in temperate rainforests in general. It should not be surprising then, that by creating small canopy openings, similar in size to what occurs in old-growth forests one can enhance habitat diversity following homogeneous disturbances such as clearcut logging. With continued monitoring, and experimentation with different ways of applying gap treatments on the landscape it should be possible to develop management techniques to substantially increase the ecological diversity and value of second growth landscapes on Prince of Wales and elsewhere in the region.

27 Cited references

Alaback, P. B. 1982. Dynamics of understory biomass in Sitka spruce-western hemlock forests of southeast Alaska. Ecology, 63, 1932-1948.

Alaback, P. B. 1984. Plant succession following logging in the Sitka spruce-western hemlock forests of southeast Alaska: Implications for management (Gen. Tech. Rep. No. PNW-173). USDA Forest Service, Pac. NW. Forest and Range Exp. Sta. Portland, OR. 26 p.

Alaback,P.B. 1986. Biomass equations for understory vegetation in coastal Alaska: The effects of species and sampling design on biomass estimates. Northwest Sci. 60:90-103.

Alaback, P.B. 1987. Biomass-dimension relationships of understory vegetation in relation to site and stand age. p 141-148 In Wharton,Eric H.; Cunia, Tiberius. Estimating tree biomass regressions and their error. Proc. of the Workshop. May 26-30 1986. Syracuse, NY. NE-GTR-117. Broomall, PA. USDA Forest Service Gen. Tech. Rep. NE-117. 303p.

Dalgaard,P. 2002. Introductory Statistics with R. Springer. New York.

Demeo, T. 1990. Unpublished. Study plan on file, Tongass National Forest, Thorne Bay Ranger District. Dated 1990.

DeMeo, T. D. Johnson, and C. Crocker-Bedford. 1990. Short-term vegetation response and deer use of artificial canopy gaps in southeast Alaska. USDA Forest Service, Ketchikan, AK. Unpublished.

DeMeo, T., J. Martin, R.A. West. 1992. Forest Plant Association Management Guide, Ketchikan Area, Tongass National Forest. USDA Forest Service Alaska Region R10-MB-210. Juneau, Alaska.

Farr, W.H. 1984. Site Index and Height Growth Curves for Unmanaged Even-Aged Stands of Western Hemlock and Sitka Spruce in Southeast Alaska. USDA Forest Service, Pacific Northwest Research Station Research Paper PNW-326. Portland.

Hanley, T.A., Brady, W.W., 1997. Understory species composition and production in old-growth western hemlock–Sitka spruce forests of southeastern Alaska. Can. J. Bot. 75, 574–580.

Hanley, T.A. and J.J. Rogers. 1989. Estimating carrying capacity with simultaneous nutritional constraints. Res. Note PNW-RN-485. USDA Forest Service, Pacific Northwest Research Station, Portland, OR. 25 pp.

28 Hanley, T. A., C. T. Robbins and D. E. Spalinger 1989. Forest habitats and the nutritional ecology of Sitka black-tailed deer: a research synthesis with implications for forest management. USDA Forest Service Pacific Northwest Research Station, Portland, OR. Gen. Tech. Rep. PNW-230.

Hanley, T. A. 2005. Potential management of young-growth stands for understory vegetation and wildlife habitat in southeastern Alaska. Landscape and Urban Planning 72: 95-112.

Knotts, L. and M. Brown. 1995. Short-term vegetation response in artificial canopy gaps on the Craig Ranger District, Tongass National Forest. USDA Forest Service, Unpublished report, dated 6 April 1995.

McClellan,M.H. 2005. Recent research on the management of hemlock-spruce forests in southeast Alaska for multiple values. Landscape and Urban Planning. 72: 65-78.

Oliver, C. D., & Larson, B. C. 1990. Forest stand dynamics. New York: McGraw Hill

Ott, R.A. and G.P. Juday. 2002. Canopy gap characteristics and their implications for management in the temperate rainforests of southeast Alaska. Forest Ecology and Management 159:271-291.

Pickett, S. T. A. and P. S. White (1985). The ecology of natural disturbance and patch dynamics. New York, Academic Press.

Veblen,T.T. and P.B. Alaback. 1995. A comparative review of forest dynamics and disturbance in the temperate rainforests in North and South America. pp. 173-213 In: R. Lawford P. Alaback, and E.R. Fuentes(eds.). High latitude rain forests and associated ecosystems of the west coast of the Americas: Climate, hydrology, ecology and conservation. Ecological Studies Vol. 116. Springer-Verlag.

Wallmo,O.C. and J.W. Schoen. 1980. Response of deer to secondary forest succession in southeast Alaska. Forest Science 26:448-462.

Watt, A. S. (1947). "Pattern and process in the plant community." J Ecology 35: 1-12.

29 Appendix I. Background on TONGASS GAP, a spreadsheet model for estimating the effects of canopy gap treatments on stand-level deer habitat.

Purpose. This simple spreadsheet model was developed in cooperation with TNC and Tongass National Forest to provide land managers and restoration practioners with a simple means to estimate the potential maximum effects of varying the number, spacing and size of canopy gap treatments on overall deer carrying capacity of a given closed-canopy second growth stand. Results provide an estimate of maximum potential deer carrying capacity for either winter or summer growing seasons derived from the initial results of a 20 year remeasurement of canopy gap treatments on Prince of Wales Island.

Assumptions: This model only considers the amount of vegetation that would be expected to develop on a well-drained moderate to highly productive Sitka spruce-western hemlock forest that was clearcut logged 45 years ago, and was treated at a stand age of 20 years. Note that this calculator can be used for either thinned or unthinned stands since in this study no significant difference was detected in vegetation response between these treatments 20 years after treatment. No consideration is made for the wide range of additional factors that could potentially effect the size of a deer population, such as winter snowpack, predation by wolves, habitat fragmentation, human hunting pressure, road use, or extreme weather events. Estimates of deer carrying capacity were derived from the model FRESH developed by Tom Hanley and colleagues in which plant biomass estimates from the gap monitoring study were used to estimate deer days for does with a fawn in the summer or adults in the winter (http://cervid.uaa.alaska.edu/deer/, Hanley and Rogers 1989, Hanley et al. 1989).

How to Use: The user enters the size of a stand in acres, then the diameter of gaps to be installed (in feet), and the number of gaps that will be installed. TONGASS GAP will then compute the proportion of the stand which will have gap habitats, the average spacing of the gaps assuming they are evenly spaced across the stand, and the maximum number of deer that could be supported in a 120 day season in the winter or the summer. Alternatively you can enter the stand size and the percentage of gap habitat to be created and it will estimate deer carrying capacity in summer and winter. This could be useful for setting habitat goals, after which you can then try various combinations of gap size and density to get the desired proportion of gap habitat. You can also change the carrying capacity estimates directly for thinned or gapped habitats in summer or winter if you have site specific data.

30 Calculations

1. Gap layout. To calculate gap area and average spacing between gaps standard geometric formulae are used.

2. Deer carrying capacity. Values are calculated from either the user entered % of stand area in gaps, or if this column is left blank it is calculated from the number and size of gaps, and the size of the stand input by the user. A weighted average is then calculated for deer carrying capacity between the proportion of the stand with gap treatments and without these treatments.

3. Timber volume reductions. This is simply based on the reduction in untreated area that results from gap treatment. This is probably an over-estimate of reductions in timber volume since no allowance has been made for gap creation resulting in greater side lighting and increases in growth rate of trees adjacent to the gap.

General Guidelines for Gap Prescriptions Derived from this Simple Calculator

Table 1. Proportion of gap habitat needed for a given deer carrying capacity goal for a hypothetical 100 acre stand. Percent Gap Summer Deer Winter Deer Habitat Carrying Capacity Carrying Capacity 0 8.3 4.17 5 9.3 5.21 10 10.2 6.25 15 11.1 7.29 20 12.0 8.33 25 12.9 9.38 30 13.8 10.42 35 14.8 11.46 40 15.7 12.5 45 16.6 13.54 50 17.5 14.58

31 Table 2. Examples of gap size, density and spacing to achieve management goals with 60’ diameter gaps in a hypothetical 100 acre stand. Number of Gaps Proportion of Average Gap Summer Deer Winter Deer Gap Habitat Spacing, Feet 10 0.6 419.2 8.5 4.3 20 1.3 296.4 8.6 4.4 40 2.6 209.6 8.8 4.7 80 5.2 148.2 9.3 5.3 100 6.5 132.6 9.5 5.5 200 13.0 93.7 10.7 6.9 400 26.1 66.3 13.1 9.6 800 52.2 46.9 17.9 15.0

32